Following a proaction paradigm, 670 Ss were trained with two lists containing five of the secondary associates (R2) to Russell-Jenkins stimulus words, and were tested 2 min., 20 mm., 2 days, or 28 days later for retention of the second list. During recall, half of the Ss were administered the five stimulus words corresponding to the five R2 words of the second list (stimulated condition), and half did not receive the stimulus words (not-stimulated condition). The stimulus words were divisible into three categories, in effect varying the cultural associative probabilities at each retention interval. Altogether, there were 24 groups completing a 4 × 2 × 3 factorial design. The retention of R2s decreased with time and the effect of stimulation was to raise their level of production above groups not so stimulated. As R2s decreased with time, intrusions of R1s (primaries) and Ra-nS (sum of Ra to R n) became more numerous where the cultural probabilities suggested this ought to happen. After 28 days, Ss still showed strong evidence of the training exposure but performance was more like that of free-associating Ss than that of shorter retention groups. In the not-stimulated condition, intrusions from unidentified sources (classified as Remainder) were more numerous the longer the retention interval. Collectively, these data support the conclusion that the amount of proactive interference via specific pre-experimental word-word habits increases as a function of time. An item analysis suggested a monotonic pattern of rs for forgetting under stimulated conditions, but not under conditions of free recall. This was interpreted to mean that forgetting when stimulated was more a process of complication than simplification and resembles a process sometimes found in motor-skills retention. Other correlational analyses proved useful tools for describing forgetting; questions pertaining to the behavior of items were quite as intriguing as those about Ss. Also, more variance could be accounted for after long than short retention intervals.
An experimental word recognition system and its evaluation.
